Examining the transfer of soils to clothing materials: Implications for forensic investigations

Forensic Science International 305 (2019) 110030

Contents lists available at ScienceDirect

Forensic Science International journal homepage: www.elsevier.com/locate/forsciint

Examining the transfer of soils to clothing materials: Implications for forensic investigations Frances A. Proctera,* , Graeme T. Swindlesa,b , Natasha L.M. Barlowc a b c

School of Geography, University of Leeds, UK Ottawa-Carleton Geoscience Centre and Department of Earth Sciences, Carleton University, Ottawa, Ontario, Canada School of Earth and Environment, University of Leeds, UK

A R T I C L E I N F O

A B S T R A C T

Article history: Received 22 July 2019 Received in revised form 9 October 2019 Accepted 27 October 2019 Available online 1 November 2019

Soil forensics has proven instrumental in assisting criminal investigation, and there is an increasing demand for experimental studies on such trace evidence. Here we present the first detailed study on the influence of clothing materials on soil transfer. We adopt an experimental approach to test the transfer of five common UK soils to five different clothing materials. Our experiment is designed to represent victim or perpetrator contact with soil at the scene of a crime. We highlight the complex relationship between soil transfer and clothing material type. Whilst over half of our soils tested displayed differential transfer to different clothing materials, soil moisture content and soil type were found to have a greater influence on the transfer of soils overall. Soil transfer is typically more effective across all material types when soils are wet and saturated. However, we find the relationship between soil transfer and material type to be more complex when soils are dry, with a significant bias in soil transfer to fleece material, which we attribute to static attraction. Encouragingly, for the analysis of forensic soils recovered from clothing artefacts, each of the transfer experiments we conducted led to soil transfer to every tested material. We suggest that future empirical studies now focus on the persistence of soils over time to clothing materials after transfer has occurred, and the transfer and persistence of soil palynomorphs present within soils. © 2019 Elsevier B.V. All rights reserved.

Keywords: Soil Forensics Transfer Trace evidence Clothing

1. Introduction Forensic soil analysis has long been implemented in multiple high profile criminal justice cases, from murders such as the Soham Murders, UK, to international justice cases such as the Bosnian genocide [1,2]. In forensic scenarios, soils are frequently recovered from artefacts and clothing that may be used to link or eliminate potential scenes of crime from suspects and victims [3– 7]. The assumption of being able to utilise such evidence is underpinned by Locard’s Exchange Principle [8] which states that wherever there is contact between two objects, a ‘mutual exchange of matter’ will occur. In the specific case of clothing, trace soil evidence including pollen and mineralogical profiles have been recovered from clothing and shoes, even after cleaning in a washing machine and dry cleaning [6,9]. This highlights the importance of understanding soil as a forensic trace evidence, and its relationship with evidentiary materials such as clothing.

* Corresponding author at: School of Geography, University of Leeds, LS2 9JT, UK. E-mail addresses: [email protected] (F.A. Procter), [email protected] (G.T. Swindles), [email protected] (N.L.M. Barlow). http://dx.doi.org/10.1016/j.forsciint.2019.110030 0379-0738/© 2019 Elsevier B.V. All rights reserved.

Soils recovered from clothing materials are useful to forensic investigations as they can be used in both investigative intelligence, for instance in narrowing search areas for crime scenes and/ or suspects [10], and also in court evidence to provide potential exclusionary evidence. It is the heterogeneity of soils which makes them so useful to criminal investigations; each soil is created from and characterised by unique geological and biological origins. Therefore, soils are extremely useful trace evidence, and can be subjected to multiple analyses of both the organic and inorganic fractions. From the inorganic soil fraction, colour [11,12], particle size distribution [13–15], mineral composition [3,16–19], elemental composition analysis [3,7,20], particle size distribution [21] and scanning electron microscopy [22–24] can all be used to differentiate between forensic soils. In terms of the organic fraction, soil organic matter [25], mycology [26] and palynology [26–28], can prove to be powerful tools in forensic science as no two locations’ palynological profiles have yet been found to be precisely identical [26]. Research suggests the characteristics of the ideal trace evidence to be: (1) that traces can be almost invisible; (2) traces are highly individualistic; (3) there is a high probability of transfer and retention; (4) traces can be quickly collected, separated and concentrated; (5) smallest traces are easily characterised; and

2

F.A. Procter et al. / Forensic Science International 305 (2019) 110030

(6) there is capacity for a computerised database from which to search [29,30]; all of which may suitably be applied to describe forensic soils. Whilst the key characteristic of ‘a high probability of transfer and retention’ is relatively well supported in published case material, it is yet to be empirically tested for soil on clothing materials. There is great abundance of excelling forensic soil science being applied to true-crime cases documented in the literature [3,4,31], however the need for further empirical research continues to prevail. In a 2017 report by the UK Forensic Science Regulator [48], it is explicitly recommended that forensic science research now focusses on ‘structured studies on the transfer and persistence of trace evidence and the significant factors affecting such transfer’. While there has been much recent interest in transfer and persistence studies of trace evidence [32–39], there have been no studies focussing on how different clothing materials affect the transfer of soil. A recent attempt to understand soil transfer patterns onto nylon-elastane bras [40,41] found mineralogy, clay content and moisture content especially to influence soil transfer. This research suggests that multiple factors affect the transfer of soil, however the influence of clothing materials is yet to be tested. As such, we present the first study towards understanding the transfer of soils to common clothing materials and the factors which may affect such transfer. 1.1. Aims In this paper, we aim to identify how soils transfer to different common clothing materials, by conducting an extensive laboratory-based soil transfer experiment. Specifically, we aim to quantify the amount of soil transferred to five different clothing materials, exposed to the same conditions, and to determine the relative influence of clothing material type on soil transfer under controlled laboratory conditions. 1.2. Hypotheses Multiple studies have been conducted on the transfer and persistence of artificial trace evidence, such as glass fragments, textile fibres, gunshot residue and fluorescent powder to clothing materials, which suggest that clothing material type does influence the transfer and persistence of trace matter [9,42,43]. As such we test the following hypotheses: 1) clothing material will influence the quantity of soil transferred; and 2) open-weaved clothing materials (cotton and denim) and fibrous material (fleece) will enable a greater amount of soil transfer than smoother and closedweave materials (nylon and leatherette). 2. Methods and materials We conducted a laboratory-based soil transfer experiment with 225 individual runs (including replicates), to determine the extent to which five UK soils transferred to five commonly worn clothing materials. We constructed a scenario designed to imitate brief contact of a person’s clothing with soil. 2.1. Soil samples Five different soil types were collected within a 30 km radius across West Yorkshire, UK (Figs. 1 and 2; Table 1). All soils are from the same underlying bedrock geology, which are sandstones of the Carboniferous Millstone Grit Group. A woodland soil was collected from a semi-rural woodland in Todmorden; and a parkland soil was collected from the nearby semi-rural public park. A streamside alluvial soil was collected from an accessible river bank behind a suburban recreation park, Todmorden. An organic-rich peaty soil

was collected from Ilkley Moor, an upland moorland environment. A glacial till-derived soil was also collected from Ilkley Moor, accessed via a river cutting. The protocol we used for soil sample collection is outlined in Fig. 3. Table 1 presents characteristics of each soil including basic descriptions, colour, pH, organic content, and particle size distribution. 2.2. Clothing materials Five commonly worn clothing materials: (1) cotton; (2) denim; (3) fleece; (4) nylon; (5) and leatherette (faux leather PVC), were selected and purchased new from a textiles shop (Table 2). Materials were used new and unwashed. To simulate a clothed body, materials were cut into swatches approximately 400
400 mm to wrap around a custom 2 kg weight (Fig. 4). Cotton, denim, fleece and nylon swatches were secured to the weight with elastic bands, and the less flexible leatherette was carefully secured with masking tape, to enable easy removal of the material from the weight after the transfer experiment with minimal impact on the soil retained. 2.3. Soil transfer experiments To create a controlled and repeatable experiment, we implemented the successful and field-verified laboratory set up design of Murray et al. [40,41] with minor modifications (Figs. 3 and 4). The different fabric swatches were wrapped around a custom 2 kg weight providing a 116.89 cm2 area of contact (Fig. 4), and were dragged over the soils for 5 s. The simulated soil surface was a 3.7 L Perspex tray filled with bulk soil, with sharp clasts, twigs >20 mm, and pebbles >20 mm removed to prevent damage to the clothing materials. The soil was levelled but not compacted. The cladded weight was placed on the soil and dragged from one side of the soil surface to the other, for a duration of five seconds, and a total contact duration of 15 s (Fig. 3). After each soil transfer experiment the clothing material was photographed whilst attached to the weight. Photographs were taken on a Canon PowerShot SX40 14 megapixel digital camera, of similar specification typically available to a forensic investigator at the scene of a crime. Photomicrographs were taken using an APEX MiniGrab USB microscope camera and Brunel light microscope, to enable a closer look at the fabric-soil interactions. Fabric swatches were then removed from the 2 kg weight and weighed to 4 decimal places. Elastic bands were cut to remove the materials and to reduce disturbance of any transferred soil. Each soil transfer experiment was run in triplicate. 2.4. Soil moisture It is suggested that soil moisture has a significant influence on soil transfer [40,41], therefore we conducted our soil transfer experiments at three different moisture contents: ‘dry’, ‘wet’, and ‘saturated’. Each experiment was first conducted with wet soil, akin to field moisture. Wet soils varied in moisture content between 25.4% and 57.7%, however were all visibly wet, and wet to the touch. Deionised water was then added to the soils until the point of saturation and the experiment repeated. Due to the nature of the different soils, the wetness and saturation point was different for each soil (Table 3). Throughout wet and saturated transfer experiments, a mist spray of deionised water was used to counteract evaporation and maintain constant moisture content. Soils were then air dried before conducting the experiment ‘dry’. When altering the moisture content of the soils, a Delta-T theta soil-moisture probe was used to estimate soil moisture content to allow samples to be as comparable as possible. Absolute moisture contents were determined by oven drying subsamples at 105  C (Table 3).

F.A. Procter et al. / Forensic Science International 305 (2019) 110030

3

Fig. 1. Location map of soil sample collection sites: A) semi-rural public parkland soil: 53 710 77. 6500 N, 2 100 60.32700 W; B) woodland soil: 53 710 59.6800 N, 2 100 60.28200 W; C) streamside alluvial soil: 53 700 51.7100 N, 2 110 81.197; D) Ilkley Moor, moorland peaty soil: 53 910 44.9300 N, 1 800 66.187 W; E) Ilkley Moor, glacial till-derived soil: 53 910 62.9500 N, 1 800 96.55300 W. Image from Google Maps 2019 (viewed 1 July 2019).

2.5. Post-transfer analysis Image analysis software imageJ (1.39 u – documentation and downloads at website http://rsbweb.nih.gov/ij/, National Institutes of Health, Bethesda, Maryland, USA), was used to quantitatively determine the amount of soil transferred to each material by calculating the total percentage cover of soil on the fabric (Fig. 3). Adjusting the image threshold highlighted the soil from the background clothing material (Fig. 5). Pixels containing transferred soil were counted, enabling us to calculate the total area of soilstained material, as a percentage of the total area of material in contact with the soil. This provided us with a comparable set of soil transfer values for each clothing material. The amount of soil transferred was also recorded by weighing the clothing material swatches pre- and post-soil transfer. Transfer weight data was corroborated against the image analysis to further examine the nature of soil transfer to the different clothing materials. Statistical

analyses were conducted on the more robust image analysis data to quantify any differences. 3. Results The results of our soil transfer experiments are presented in Fig. 6. It is clear there is great variation in soil transfer, with a greater quantity of soil transfer with increasing soil moisture content. As such, to enable a better insight into the relative influence of clothing materials specifically on soil transfer, we make our comparisons within the three moisture content subsets. Soil transfer measured in percentage area of soil coverage and weight of soil transfer are available in the supplementary information. ANOVA tests for differences conducted between the mean soil transferred (area coverage) to the different clothing materials identify significant differences for nine of the fifteen soils tested

4

F.A. Procter et al. / Forensic Science International 305 (2019) 110030

Fig. 2. Site photographs of soil sample collection: A) parkland soil; B) woodland soil; C) streamside alluvial soil; D) moorland peaty soil; E) glacial till-derived soil.

F.A. Procter et al. / Forensic Science International 305 (2019) 110030

5

Table 1 Characteristics of the five soil types used in this study including: soil colour; pH; organic matter; and particle size distribution of the <2 mm fraction where: coarse sand (2 mm-500 mm); medium sand (500 mm-150 mm); fine sand (150-64 mm); and silt and clay (63 mm). Soil type

Colour

pH

Organic content

Particle size distribution

Description

Parkland soil

2.5Y 4/2 Dark greyish brown

6.06

9%

Sandy loam. 10% mottling of sand grains. Anthropogenic inclusions of brick, pottery and plastics <1%. Modified anthroposol.

Woodland soil

5Y 2.5/2 Black

5.47

33%

Moorland peaty soil

5YR 2.5/1 Black

3.52

50%

Streamside alluvial soil

10YR 4/4 Dark yellowish brown

4.80

9%

Glacial till-derived soil

2.5Y 5/6 Light olive brown

3.61

4%

Coarse sand: 17% Medium sand: 68% Fine sand: 11% Silt and clay: 4% Coarse sand: 62% Medium sand: 26% Fine sand: 9% Silt and clay: 3% Coarse sand: 53% Medium sand: 31% Fine sand: 11% Silt and clay: 5% Coarse sand: 26% Medium sand: 56% Fine sand: 13% Silt and clay: 5% Coarse sand: 46% Medium sand: 38% Fine sand: 9% Silt and clay: 7%

Sandy loam. 5% large organic inclusions, twigs and seeds

Peaty soil. High organic content

Sandy sediment

Very clay rich till, with a sticky texture. Clast inclusions 20%

Fig. 3. The protocol used in this soil transfer experiment, including soil collection, laboratory soil transfer, and post-transfer analysis in imageJ.

6

F.A. Procter et al. / Forensic Science International 305 (2019) 110030

Table 2 Photographs (1x magnification), photomicrographs (scale bar is 500 mm) and properties of the five clothing materials used in this study. Photograph

Material Composition

Fibre Type

Characteristics

Cotton

Photomicrograph

100% cotton

Natural

Coarse open weave and fibrous

Denim

100% cotton

Natural

Coarse open weave and fibrous

Fleece

100% acrylic polyester

Synthetic

Artificial fibrous texture

Nylon

100% nylon

Synthetic

Closed weave, square textured pattern

Leatherette

100% PVC

Synthetic

Closed weave, smooth texture with some mottled pattern

(Supplementary Table 1). Both cotton and fleece materials consistently prevail in having the greatest mean soil transfer. By fitting linear regression models, we are able to further statistically explore the relationships between soil transfer to the different materials and the relative influence of the factors affecting this transfer. We find that soil moisture has the greatest influence on soil transfer to clothing materials. The linear regression model between soil transfer and soil moisture (R2 = 0.782, p = 0.000) reflects a strong significant relationship between soil moisture content and the amount of soil transferred. Our linear regression model between soil transfer, soil moisture, soil type, and clothing material type, however (R2 = 0.809, p = 0.000) highlights that soil type and fabric type do contribute to the degree of soil transfer. The extent to which clothing material type and soil type affects soil transfer of dry, wet and saturated soils is described below. 3.1. Dry soils In our dry soils, clothing material type has a greater influence on soil transfer than soil type, with linear regression models of R2

= 0.515, p = 0.000 and R2 = 0.000, p = 0.110 respectively. Dry soils demonstrate the greatest variability in soil transfer between clothing materials. Transfer of each of the dry soils is greater to cotton and denim (natural and open weaved fabrics), than the nylon and leather (both synthetic and closed weaved). The greatest quantity of soil transfer is however to the fleece material, with a notable increase in soil transfer of all the dry soils to this material. This significant transfer is highlighted in Fig. 7. We attribute this predominantly to the static nature of the fabric. This bias in transfer to fleece material is also typically reflected in the weight data. Of the dry soil transfers weighed, only the glacial-till derived soil was not most abundant in transfer to fleece – where the claylike material may not have been attracted statically. The driest soil sample, the dry streamside alluvial soil (only 2% moisture content) proved difficult to calculate a percentage-area coverage and so was excluded from image analysis. Due to the extremely low moisture content and lack of compaction the dry alluvial soil created a fine dust, which covered the fabrics (Fig. 8). This fine dusting was assumed to be the background material colour by the image analysis software, with only larger soil clasts being detected. The

F.A. Procter et al. / Forensic Science International 305 (2019) 110030

7

Fig. 4. Experiment laboratory set up: A) 2 x 1 kg weights secured together; B) 2 kg weight and strong plastic pull-line; C) cladded weight in denim secured with an elastic band, being dragged across the glacial till-derived soil surface; D) laboratory workspace showing the soil surface set up, in-situ photography stand, and weighing balance.

weight of soil transfer here is however still most abundant to fleece, followed by the open weaved cotton and denim, and least transferred to the closed weaved nylon and leatherette. Across all of the experiments, the lowest recorded soil transfer is 3.88 mg of dry glacial till-derived soil transferred to the leatherette. Although not statistically significant, there is a notable degree of variation in soil transfer between the different dry soil types, both in area of soil coverage and weight of transfer (Supplementary information). Table 3 Soil moisture content (%) for each soil in the transfer experiment.

Parkland soil Woodland soil Streamside alluvial soil Moorland peaty soil Glacial till-derived soil

Dry

Wet

Saturated

23.0 40.8 1.9 24.5 11.4

31.3 57.7 21.7 55.3 25.4

34.0 64.0 26.4 63.1 54.1

3.2. Wet soils Within the wet soil transfer experiments, the linear regression between soil transfer, material type and soil type shows a significant relationship R2 = 0.654, p = 0.000. Contradictory to the drier soils, soil type influences the amount of soil transfer to a greater extent than clothing material type with linear regression models of R2 = 0.461, p = 0.000 and R2 = 0.084, p = 0.003 respectively. Compared with the dry transfer experiments, transfer is more ubiquitous between the materials. Cotton consistently yields a high percentage cover of soil, with denim and fleece picking up slightly less material. The weight of soil transferred suggests denim and leather enable the greatest amount of wet soil transfer. Interestingly, both nylon and leather display relatively high amounts of soil transfer, despite having a smoother finish similar to waterproof-type materials. The fleece material consistently recovers relatively lower amounts of wet soil than the other

8

F.A. Procter et al. / Forensic Science International 305 (2019) 110030

Fig. 5. Adjusted image threshold in imageJ highlighting only the soil on the photograph, and area of interest highlighted using the oval tool, next to the original photograph.

Fig. 6. Average percentage area of soil coverage on each clothing material, following soil transfer. Dry streamside alluvial soil omitted.

fabrics, both in percentage cover of soil staining and in weight of transfer. 3.3. Saturated soils The influence of both clothing material and soil type is less significant in more saturated soils (R2 = 0.474, p = 0.000). Again, soil

type appears to have a greater relative influence on soil transfer than material type, with liner regression models R2 = 0.259, p = 0.000 and R2 = 0.126, p = 0.001 respectively. Transfer of these saturated soils again appears more ubiquitous across the different materials. Unlike the dry soils, fleece appears to pick up the least amount of saturated soils. This is however still a great amount of soil transfer with the least transfer still covering approximately

F.A. Procter et al. / Forensic Science International 305 (2019) 110030

9

Fig. 7. Transfer of dry woodland soil to A) cotton; and B) fleece. Scale bar is 20 mm.

Fig. 8. Transfer of dry streamside alluvial soil to A) nylon; and B) leatherette. Scale bar is 20 mm.

50% surface area of the material. While there appears to be great variation in the weight of soil transferred to each material, the least amount of saturated soil transfer, the saturated glacial till-derived soil transferred to fleece, still transferred a great amount of soil; weighing 1804.9 mg. In summary, soil moisture is the dominant influence on soil transfer, however soil type and clothing material type do also contribute to the amount of soil collected onto the material. The relative influence of the clothing material decreases with increasing soil moisture, with material type being most influential on the transfer of drier soils. Transfer of soil is more ubiquitous across fabric types tested when soils are wet. 4. Discussion 4.1. Explanation of results We present the first empirical study investigating the relative influence of clothing material on the transfer of soil. We hypothesised that material type will affect the amount of soil transferred in this controlled experiment, and specifically that more open weaved and fibrous materials would enable a greater amount of soil transfer. Our results show that the relationship between soil transfer and clothing material is existent, yet complicated. In wetter soils, where transfer of soils is greater, the relative influence of clothing material is reduced, and soil type

becomes a relatively more influential factor. In drier soils however, clothing material can play a significant role in determining the degree of soil transfer. The greatest amount of transfer occurred between the glacialtill derived soil and the leatherette material. In the wettest and stickiest soils, suction occurred between the soil and the materials, and larger clumps of soil particles were observed to have transferred. The highest recorded weights of soil transfer occur where large aggregates of soil particles stuck to the fabric, the volume of which is not necessarily detected in the image analysis. The lowest recorded soil transfer in the experiment: 3.88 mg of dry glacial till-derived soil, is extremely encouraging for the forensic analysis of soil recovered from clothing as it is enough to conduct many soil analysis techniques, including analysis of SOM, polarised light microscopy, and potentially XRD. The most notable pattern observed was the significantly higher transfer of dry soils to fleece than to any other material. We attribute this to the static nature of the material, which was clearly observed in the laboratory. This static attraction enabled more dry soil particles to be lifted and adhered to the material. There is little discussion in the forensic and trace evidence literature documenting the influence of static on particulate transfer. The adhesion of ‘dirt’ to fleece by static attraction is however documented; in a patent design for new fleece with less static property ([44], U.S. Patent 7,213,313). This may prove to be a potential field of trace evidence research in itself.

10

F.A. Procter et al. / Forensic Science International 305 (2019) 110030

We hypothesised that materials with a more open weave will demonstrate a greater level of soil transfer, as rougher surfaces tend to entrain larger numbers of particles. Multiple studies on the transfer of inorganic materials found a coarser, and more open weave to allow a greater degree of particulate retention [42,43]. Our photomicrographs show particles, and some larger particles entrained in these open weaved fabrics (Fig. 9). Particles did transfer surprisingly well to the closed weaved nylon and leather, however both had some degree of texture which retained soil particles (Fig. 9). The persistence of soils to these materials would prove an interesting study through future experimentation. Where there is significant variation within material groups, we suggest some variation could be explained by the weighted clothing material not being completely flat on the soil surface, however the soil surface was relevelled after each experiment, so variations are representative of real-life variation. 4.2. Implications of findings We find some of our most interesting results in the transfer of dry soils. One of the crucial characteristics of trace evidence is its

potential to go unnoticed by a perpetrator [29]. With smaller quantities of soil transferred when dry, a suspect may be less aware of the presence of this trace evidence on their clothing, compared with the saturated soils where staining is more obvious. In the lowest yielding soil transfer experiment enough soil was still present to allow for subsequent forensic analyses, which reflects the crucial need to further understand the transfer of soil trace evidence to clothing materials. Even in the lowest yielding soil transfers, or where recovery of soil may prove difficult, the colour of the soil remained well represented on the materials, which can prove to be critical evidence alone with variations in soil colour one of the most distinguishing characteristics of trace soil evidence [11,12]. From our increasing understanding of the dynamics of soil transfer we may enable better interpretation of the trace evidence and can also improve approaches for the recovery of such evidence material. Understanding differential soil transfer to clothing materials may allow us to make future recommendations on the methods by which soil is recovered from evidential materials. For instance, if dry soil is found adhered to nylon or leather, where transfer is least, it may be most appropriate to use the direct tag

Fig. 9. Photomicrographs showing closer detail soil-material interactions: A) woodland soil embedded in the weave of cotton; B) glacial till-derived soil embedded and in the fibres of denim; C) large woodland soil fragments intertwined in fleece fibres; D) parkland soil transferred along the pattern of textured nylon; E) streamside alluvial soil scraped across leatherette, adhering predominantly on the raised texture of the material. Scale bar is 1 mm long.

F.A. Procter et al. / Forensic Science International 305 (2019) 110030

lifting method [24] to yield evidential soil for analysis; whereas on wetter soils or materials which yielded a high degree of soil transfer, the standard washing protocol to recover the soil should prove sufficient [24,45]. We suggest the need to create a database for all empirical research studies on soil forensics, which may act as a reference base for casework. Here, we have 225 stock images of soil transfer, when five of the most commonly worn clothing materials come into contact with common soils across the UK. The potential of such a database is reflected in the demonstration of how image processing of crime scene photographs can provide initial and potentially crucial evidence on the colour, nature, and subsequently the provenance of soil trace evidence in a non-destructive manner [40,41]. From such stock evidence, forensic investigators may be able to determine valuable information such as retracing potential antecedent weather conditions or ground conditions at the time of incident, narrowing a window of opportunity when such evidence may have come to be. This would also be beneficial to investigators who are not geologically trained to better understand the uniqueness of soil evidence, and to enable better consideration for the potential origins of such evidence. The creation of such databases would only strengthen the criterion of soils as an ideal trace evidence [29]. This research may also have implications for biosecurity research. A great amount of research is currently focussing on the control of invasive species, the minimum amount of trace soil and plant material that can contribute to invasive species spreading, and best practice when working in the field. Our findings may prove useful when considering the most appropriate field attire when working on and around invasive species. 4.3. Limitations Whilst the imageJ software was in the most part successful in independently determining the contrast between soil and the background fabric, in multiple instances with the denim swatches in particularly the image threshold needed to be adjusted manually, with the user determining when the software had picked up all of the transferred soil and excluded background noise. Whilst this may account for a slight margin-of-error, due diligence was taken by adjusting the threshold side-by-side with the original transfer photographs to allow the best possible estimation of soil transferred. We recognise a potential limitation to this study being the replication of field activity in the laboratory, with not all natural factors being successfully replicated, for instance soil compaction. This laboratory study is however designed based on a field-verified method [40,41], and did also allow for better control of other factors, such as soil moisture content, levelling, and better crosscontamination control. This experiment could be repeated in the field, with the provision of high precision field-portable scales and contamination measures in place. 4.4. Future work and recommendations The geographical scope of this study has the potential to be extended considerably. Whilst chosen to be representative of common UK soil types, our five soils tested do not represent all UK soils, and it is the heterogeneity of soils which makes them so useful to forensic investigations. We reinforce the suggestion that a database of all studies be compiled, to further enhance our understanding of how this unique trace evidence behaves, and to build a better picture of how soil evidence may be useful in future forensic investigation. We have quantified the amount of soil transferred to multiple common clothing materials in a forensic setting, however there is

11

suggestion that a rapid decay rate of transference can quickly obliterate valuable trace evidence on some fabric surfaces [46]. Reinforcing the recommended research focus from the UK Forensic Science Regulator, we suggest a follow-up study be conducted to determine the persistence of such transferred soil traces on the different clothing materials over time. One requisite of the ‘ideal trace evidence’ is that the evidence should be ‘nearly invisible’ [29]; however, in our experiment quite a lot of material was often visible following transfer. Denim is the only non-white material used in this experiment and soil staining was less obvious on denim, especially for the drier soils. Soil transfer may become increasingly less obvious when transferred to darker materials. If soil from a forensic scenario did result in a larger and more noticeable soil stain such as those of the saturated soils, one could assume that any perpetrator would aim to remove such a stain. It is for this reason that considering the persistence of evidence is also crucially important. One outstanding question is how much soil would remain adhered to these materials following washing, or another form of evidence diminishment. As such, we recommend future studies should test the persistence of different soils following transfer to clothing materials under different treatments to remove material. Where soil transfer is ‘nearly invisible’, or present in very small traces (e.g. following transfer of dry soil), consideration must also be given to any background traces already present on a material. There is a suggestion that the often non-sterile environments of production of some materials, can contaminate materials to an extent where background traces mask any potential signal from trace evidence [47]. It is imperative that any forensic signals are separated accurately from background traces [34]. Methods such as washing, dry brushing and dry tag lifting to recover soil from stained clothing have all proven effective in recovering representative populations of particles from evidence material [24]. Having highlighted that all clothing materials tested retained at least some soil, with most transferences allowing for multiple forensic analyses to be conducted on the collected soil, we recommend future work investigates the potential that can be recovered from soil adhering to clothing, for instance the transfer and persistence of soil palynomorphs. Much research has been done already on the inorganic fraction of forensic soils, with focusses on mineral and elemental composition specifically. Soils each possess a unique signature, and palynological profiles of the lesser researched organic fraction can prove particularly useful [49,50]. It would be greatly beneficial to determine how this palynological signature also transfers to clothing materials from forensic soils. 5. Conclusion We present the first experimental investigation into the transfer of soils to different clothing materials, and the factors which influence the transfer of this forensic trace evidence. We conclude that soil moisture, and typically soil type have a greater influence on the amount of soil transfer, than clothing material type. Clothing material type does however have a significant influence on the transfer of drier soils. An unexpected finding was the extent to which static attraction influenced the transfer of drier soil particles to fleece material, which has previously been little explained in forensic trace studies. The key outputs from our empirical research are:  We find all common clothing materials tested retained trace soil after brief contact with a soil surface, which is highly encouraging for the forensic analysis of soils recovered from clothing;  Our findings confirm the important influence of soil moisture on the quantity of soil transferred during contact;

12

F.A. Procter et al. / Forensic Science International 305 (2019) 110030

 The relative influence of clothing material type on soil transfer decreases with increasing soil moisture;  225 stock images of soil transfer to common clothing materials under different soil conditions have been created;  We suggest forensic investigators may consider using different techniques to retrieve soil from different clothing materials;  We recommend further study is conducted into the transfer and persistence of soils, and the lesser researched soil palynomorphs, to clothing materials.

Declaration of Competing Interest None. CRediT authorship contribution statement Frances A. Procter: Conceptualization, Methodology, Validation, Formal analysis, Investigation, Writing - original draft, Visualization, Project administration. Graeme T. Swindles: Conceptualization, Writing - review & editing, Supervision. Natasha L.M. Barlow: Conceptualization, Writing - review & editing, Supervision. Acknowledgements This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. FP was supported in part by the University of Leeds Alumni Bursary, for her MSc research. Appendix A. Supplementary data Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.forsciint.2019.110030. References [1] A. Ruffell, Forensic geoscience, Geol. Today 22 (2) (2006) 68–70. [2] T. Kukes, Defining Genocide: A Biological Perspective of Mass Burials in Srebrenica, Bosnia-Herzegovina MSc Thesis, University of WisconsinMilwaukee, 2015. [3] N. Petraco, T.A. Kubic, N.D. Petraco, Case studies in forensic soil examinations, Forensic Sci. Int. 178 (2–3) (2008) e23–e27. [4] G.T. Swindles, A. Ruffell, A preliminary investigation into the use of testate amoebae for the discrimination of forensic soil samples, Sci. Justice 49 (3) (2009) 182–190. [5] R.W. Fitzpatrick, M.D. Raven, How pedology and mineralogy helped solve a double murder case: using forensics to inspire future generations of soil scientists, Soil Horiz. 53 (5) (2012) 14–29. [6] R. Fitzpatrick, M. Raven, P. Self, Clay mineralogy as significant evidence in 4 murder investigations involving a wide range of earth materials from Perth, Adelaide, Melbourne and Sydney, Australian Clay Minerals Society Conference–Perth (2014) 3–5. [7] S.J. Singletary, H.D. Hanna, Forensic soil analysis using the electron microprobe: the markice-bowling case, Microsc. Microanal. 24 (S1) (2018) 1182–1183. [8] E. Locard, Analyses of dust traces parts I, II and III, Am. J. Police Sci. 1 (1930) 276–298, 401–418 and 496–514. [9] P.A. Bull, R.M. Morgan, A. Sagovsky, G.A.J. Hughes, The transfer and persistence of trace particulates: experimental studies using clothing fabrics, Sci. Justice 46 (3) (2006) 185–195. [10] J.K. Pringle, A. Ruffell, J.R. Jervis, L. Donnelly, J. McKinley, J. Hansen, R. Morgan, D. Pirrie, M. Harrison, The use of geoscience methods for terrestrial forensic searches, Earth Rev. 114 (1–2) (2012) 108–123. [11] E.P. Junger, Assessing the unique characteristics of close-proximity soil samples: just how useful is soil evidence? J. Forensic Sci. 41 (1) (1996) 27–34. [12] R. Sugita, Y. Marumo, Validity of color examination for forensic soil identification, Forensic Sci. Int. 83 (3) (1996) 201–210. [13] R.J. Dudley, The particle size analysis of soils and its use in forensic science— the determination of particle size distributions within the silt and sand fractions, J. Forensic Sci. Soc. 16 (3) (1976) 219–229.

[14] S.J. Blott, D.J. Croft, K. Pye, S.E. Saye, H.E. Wilson, Particle size analysis by laser diffraction, Geol. Soc. Lond. Spec. Publ. 232 (1) (2004) 63–73. [15] K. Pye, S.J. Blott, Particle size analysis of sediments, soils and related particulate materials for forensic purposes using laser granulometry, Forensic Sci. Int. 144 (1) (2004) 19–27. [16] W. Graves, A mineralogical soil classification technique for the forensic scientist, J. Forensic Sci. 24 (2) (1979) 323–338. [17] A. Ruffell, P. Wiltshire, Conjunctive use of quantitative and qualitative X-ray diffraction analysis of soils and rocks for forensic analysis, Forensic Sci. Int. 145 (1) (2004) 13–23. [18] R.C. Murray, Evidence from the Earth: Forensic Geology and Criminal Investigation, Mountain Press Publishing, 2004. [19] A. Ruffell, J. McKinley, Forensic geoscience: applications of geology, geomorphology and geophysics to criminal investigations, Earth Sci. Rev. 69 (3–4) (2005) 235–247. [20] S. Raut, Forensic Physics [Online]. [Accessed 22 November 2018]. Available from:, (2012) . http://santoshraut.com/forensic/physics.htm. [21] R. Sugita, Y. Marumo, Screening of soil evidence by a combination of simple techniques: validity of particle size distribution, Forensic Sci. Int. 122 (2–3) (2001) 155–158. [22] S. Cengiz, A.C. Karaca, I. Çakır, H.B. Üner, A. Sevindik, SEM–EDS analysis and discrimination of forensic soil, Forensic Sci. Int. 141 (1) (2004) 33–37. [23] A. Ruffell, J. McKinley, Geoforensics, John Wiley & Sons, 2008. [24] D. Pirrie, Testing the efficiency of soil recovery from clothing for analysis by SEM-EDS, Forensic Sci. Int. 289 (2018) 83–91. [25] V.F. Melo, J.M. Mazzetto, J. Dieckow, E.J. Bonfleur, Factor analysis of organic soils for site discrimination in a forensic setting, Forensic Sci. Int. 290 (2018) 244–250. [26] P.E. Wiltshire, D.L. Hawksworth, J.A. Webb, K.J. Edwards, Palynology and mycology provide separate classes of probative evidence from the same forensic samples: a rape case from southern England, Forensic Sci. Int. 244 (2014) 186–195. [27] D.C. Mildenhall, P.E. Wiltshire, V.M. Bryant, Forensic palynology: why do it and how it works, Forensic Sci. Int. 163 (3) (2006) 163–172. [28] P.E. Wiltshire, Protocols for forensic palynology, Palynology 40 (1) (2016) 4–24. [29] K. Aardahl, Evidential Value of Glitter Particle Trace Evidence Doctoral Dissertation, National University, San Diego, 2003. [30] R.W. Fitzpatrick, M.D. Raven, S.T. Forrester, A systematic approach to soil forensics: criminal case studies involving transference from crime scene to forensic evidence, in: D. Miller, L. Dawson (Eds.), Criminal and Environmental Soil Forensics, Springer, Dordrecht, 2009, pp. 105–127. [31] A. Ruffell, B. Schneck, International case studies in forensic geology: fakes and frauds, homicides and environmental crime, Episodes: J. Int. Geosci. 40 (2) (2017) 172–175. [32] R.M. Morgan, J. Flynn, V. Sena, P.A. Bull, Experimental forensic studies of the preservation of pollen in vehicle fires, Sci. Justice 54 (2) (2014) 141– 145. [33] K.R. Scott, R.M. Morgan, V.J. Jones, N.G. Cameron, The transferability of diatoms to clothing and the methods appropriate for their collection and analysis in forensic geoscience, Forensic Sci. Int. 241 (2014) 127–137. [34] D.A. Stoney, A.M. Bowen, P.L. Stoney, Loss and replacement of small particles on the contact surfaces of footwear during successive exposures, Forensic Sci. Int. 269 (2016) 78–88. [35] E.A. Levin, R.M. Morgan, K.R. Scott, V.J. Jones, The transfer of diatoms from freshwater to footwear materials: an experimental study assessing transfer, persistence, and extraction methods for forensic reconstruction, Sci. Justice 57 (5) (2017) 349–360. [36] J.C. Webb, H.A. Brown, H. Toms, A.E. Goodenough, Differential retention of pollen grains on clothing and the effectiveness of laboratory retrieval methods in forensic settings, Forensic Sci. Int. 288 (2018) 36–45. [37] A.L. Gassner, M. Manganelli, D. Werner, D. Rhumorbarbe, M. Maitre, A. Beavis, C.P. Roux, C. Weyermann, Secondary transfer of organic gunshot residues: empirical data to assist the evaluation of three scenarios, Sci. Justice 59 (1) (2019) 58–66. [38] M. Maitre, S. Chadwick, K.P. Kirkbride, A.L. Gassner, C. Weyermann, A. Beavis, C. Roux, An investigation on the secondary transfer of organic gunshot residues, Sci. Justice 59 (3) (2019) 248–255. [39] K.R. Scott, R.M. Morgan, N.G. Cameron, V.J. Jones, Freshwater diatom transfer to clothing: spatial and temporal influences on trace evidence in forensic reconstructions, Sci. Justice 59 (3) (2019) 292–305. [40] K.R. Murray, R.W. Fitzpatrick, R.S. Bottrill, R. Berry, H. Kobus, Soil transference patterns on bras: image processing and laboratory dragging experiments, Forensic Sci. Int. 258 (2016) 88–100. [41] K.R. Murray, R.W. Fitzpatrick, R. Bottrill, H. Kobus, Patterns produced when soil is transferred to bras by placing and dragging actions: the application of digital photography and image processing to support visible observations, Forensic Sci. Int. 276 (2017) 24–40. [42] T. Hicks, R. Vanina, P. Margot, Transfer and persistence of glass fragments on garments, Sci. Justice 36 (2) (1996) 101–107. [43] T.J. Allen, K. Hoefler, S.J. Rose, The transfer of glass—part 2: a study of the transfer of glass to a person by various methods, Forensic Sci. Int. 93 (2–3) (1998) 175–193. [44] Nohara, S., Silver Ox Inc, 2007. Method of manufacturing fleece having different kinds of fibers in front and back faces. U.S. Patent 7,213,313.

F.A. Procter et al. / Forensic Science International 305 (2019) 110030 [45] A. Ruffell, A. Sandiford, Maximising trace soil evidence: an improved recovery method developed during investigation of a $26 million bank robbery, Forensic Sci. Int. 209 (1-3) (2011) e1–e7. [46] R.M. Morgan, P.A. Bull, The philosophy, nature and practice of forensic sediment analysis, Prog. Phys. Geogr. 31 (1) (2007) 43–58. [47] A. Keaney, A. Ruffell, J. McKinley, Geological trace evidence: forensic and legal perspectives, in: D. Miller, L. Dawson (Eds.), Criminal and Environmental Soil Forensics, Springer, Dordrecht, 2009, pp. 221–237. [48] Forensic Science Regulator, Annual Report: November 2016–2017 [Online] Available from: https://www.gov.uk/government/publications/

13

forensic-science-regulator-annual-report-2017. p.36. Accessed May 2019, (2019) . [49] P.E. Wiltshire, Consideration of some taphonomic variables of relevance to forensic palynological investigation in the United Kingdom, Forensic Sci. Int. 163 (3) (2006) 173–182. [50] P.E. Wiltshire, D.L. Hawksworth, J.A. Webb, K.J. Edwards, Palynology and mycology provide separate classes of probative evidence from the same forensic samples: a rape case from southern England, Forensic Sci. Int. 244 (2014) 86–195.